AM2434: CPSW

Mari Tsunoda

Part Number: AM2434

Hi Team,

My customer has a question about CPSW. The i2331 errata says the following: Due to this errata, Ethernet boot should not be used on this device.

Errata link: https://www.ti.com/lit/pdf/sprz457

They do not use ethernet boot, but they use CPSW in their application. Is this workaround needed for them?
In the workaround, it says that a warm reset is needed after POR. After issuing a warm reset, is this followed by the restarting boot process? They are using xSPI boot, so will the procedure below be needed?
1. Issue POR
2. CPU starts xSPI boot, and it reads SBL/Application binary from the external flash memory.
3. SBL or application tells CTRLMMR_MCU_RST_CTRL.SW_MCUWARMRST to issue warm reset.
4. Due to warm reset, CPU starts xSPI boot and reads SBL/Application again.
5. This time SPL or application never issue warm reset.
6. SPL or application starts initializing CPSW. (Also, the CPU needs to execute boot process twice to use CPSW correct?)
In the procedure above, SBL/application does the following correct? How can SBL/application know which type of reset was issued? (Is it the CTRLMMR_MCU_RST_SRC register?)
1. Issue warm reset when boot from POR
2. Not issue warm reset when boot from warm reset.

Best regards,

Mari

over 3 years ago

0 Prasad Jondhale over 3 years ago

TI__Mastermind 27743 points

Hi Mari-san,

1. Yes, workaround is needed when CPSW is used in application.

Basically

* no workaround for boot so Ethernet boot can't be used at all

* for run-time usage of CPSW, customers needs to use workaround.

2. Regarding the workaround implementation -

Yes, your flow is correct. We will providing the implementation in our next SDK release (SDK8.4, planned on 15th of Sept).

Sharing the workaround pseudo code for reference.

+       rst_src = readl(0x04518178); // read  CTRLMMR_RST_SRC register
+       printf("MCU_RST_SRC: 0x%x\n", rst_src);
+
+       if (rst_src == COLD_BOOT || rst_src & (SW_POR_MCU | SW_POR_MAIN)) {
+               printf("do sw_mcu_warmrst now\n");
+
+               /* unlock mmr */
+               writel(0x68ef3490, 0x04519008);
+               writel(0xd172bc5a, 0x0451900c);
+
+               /* clear RST_SRC register */
+               writel(0xFFFFFFFF, 0x04518178);
+               /* do sw_mcu_warmrst */
+               writel(0x600, 0x04518170);
+       }

JIRA id for reference - [MCUSDK-4430] CPSW: Device lockup when reading CPSW registers - Texas Instruments JIRA (ti.com)

Regards,

Prasad

0 Mari Tsunoda over 3 years ago in reply to Prasad Jondhale

TI__Expert 4920 points

Hi Prasad,

Thanks for your quick response on 1&2. I will provide the pseudocode for my customer as well.

Let me know if you can answer 3 as well!

Best regards,

Mari

0 Prasad Jondhale over 3 years ago in reply to Mari Tsunoda

TI__Mastermind 27743 points

Hi Mari-San,

I send you the mail which has very detailed errata code (with lots of comments). You can use it to share with customer.

Regards,

Prasad

0 Mari Tsunoda over 3 years ago in reply to Prasad Jondhale

TI__Expert 4920 points

Hi Prasad,

Received, thank you!

Best regards,

Mari

0 Prasad Jondhale over 3 years ago in reply to Mari Tsunoda

TI__Mastermind 27743 points

Mari-San,

Hope all your queries are resolved. I will close the thread.

Prasad

0 Mari Tsunoda over 3 years ago in reply to Prasad Jondhale

TI__Expert 4920 points

Hi Prasad,

Re-opening this thread as we received some feedback on the pseudocode. I sent you the feedback as well as a question regarding it via email. Could you take a look at it for me? Also updated the thread notes for your reference.

Best regards,

Mari

0 Prasad Jondhale over 3 years ago in reply to Mari Tsunoda

TI__Mastermind 27743 points

Hi Mari,

Sure, will have a look at your queries.

Meanwhile, can you please let me know if customer is using our SDK? Just wondering if we can share patch on MCU+ SDK which our SW team has already ready instead of customer applying it on their end.

Regards,

Prasad

0 Mari Tsunoda over 3 years ago in reply to Prasad Jondhale

TI__Expert 4920 points

Hi Prasad,

Yes, they are already waiting for the next patch so I believe they will be ok with it. But since we provided the pseudocode for their end, it would be best if we could give an answer their questions.

Best regards,

Mari

0 Prasad Jondhale over 3 years ago in reply to Mari Tsunoda

TI__Mastermind 27743 points

Hi Mari-San,

Pseudocode was only for reference and not for implementation as we don't want to spend efforts in debugging when we are already providing the implementation.

I talked to SBL expert and he will share patch by tomorrow which can directly be integrated.

Regards,

Prasad

0 Mari Tsunoda over 3 years ago in reply to Prasad Jondhale

TI__Expert 4920 points

Hi Prasad,

I understand. Can you ask the SBL expert to share the patch with me?

Best regards,

Mari

+1 Prasad Jondhale over 3 years ago in reply to Mari Tsunoda

TI__Mastermind 27743 points

Hi Mari-San,

Closing the thread as patch was shared offline and confirmed to be working on customer board.

Regards,

Prasad

0 Mari Tsunoda over 3 years ago in reply to Prasad Jondhale

TI__Expert 4920 points

Hi Prasad,

Prasad Jondhale said:
our next SDK release (SDK8.4, planned on 15th of Sept).

Are there any changes regarding release date for the SDK or is it still Sep-15?

Best regards,

Mari Tsunoda

0 Prasad Jondhale over 3 years ago in reply to Mari Tsunoda

TI__Mastermind 27743 points

Yes, it is still Sep-15.

0 Mari Tsunoda over 3 years ago in reply to Prasad Jondhale

TI__Expert 4920 points

Hi Prasad,

When will it be released as it is not out yet?

Best,

Mari

0 Prasad Jondhale over 3 years ago in reply to Mari Tsunoda

TI__Mastermind 27743 points

Hi Mari-san,

SDK release is delayed by 2 weeks and new date being tracked is 27th of Sept.

Regards,

Prasad

Arm-based microcontrollers

Arm-based microcontrollers forum

AM2434: CPSW